The Spatial Distribution And Exposure Risk Assessment Of NO₂/SO₂ Concentration Simulated By Improved LUR Model

As the problem of air pollution becomes more and more serious,the research on air pollutants is increasing gradually.Atmospheric pollutant concentration is an important index used to evaluate environmental quality.The monitoring of atmospheric pollutant includes six gas concentration indexes,NO₂ and SO₂ are pollution gases that directly or indirectly harm human health,high concentration of NO₂ and SO₂ in the atmosphere was the cause of a variety of congenital diseases and respiratory diseases.In order to effectively control NO₂ and SO₂ emissions in the atmosphere,it is necessary to study the spatial distribution of NO₂ and SO₂ concentrations.In this research,a Land Use Regression（LUR）model was established based on measured data from air monitoring stations,natural environmental factors,social and economic factors,etc.,to simulate the concentration of NO₂/SO₂ at other unknown points in space,and to obtain a high-precision spatial distribution diagram of NO₂/SO₂.In addition,high-precision population spatial distribution map and pollutant spatial distribution map were used to realize the spatial division of population exposure relative risk in Anhui province,and the exposure risk data were statistically analyzed to analyze the dynamic changes of NO₂/SO₂ population exposure relative risk in different seasons and cities,the main research content and conclusions include:（1）By analyzing the original site data of different time scales,it is proved that the change of NO₂/SO₂ concentration has great spatial variability and seasonal difference,and the high-value region occurs around December and January;The influencing factors of NO₂/SO₂ concentration were selected,and bivariate analysis and correlation analysis were conducted between two variables to analyze the correlation between the influencing factors and NO₂/SO₂ concentration.Stepwise regression analysis was used to screen the independent variables;（2）The gradient lifting tree and random forest algorithm were used to improve the traditional LUR model,to construct the regression model of NO₂/SO₂ concentration,and to evaluate the fitting effect of the model before and after the improvement.Experimental results show that The improved LUR model fitted by the random forest algorithm has the best fitting effect,In NO₂ verification set,Root Mean Square Error（RMSE）and Mean Absolute Error（MAE）are 0.859,4.466 and 3.186respectively.In SO₂ verification set,R Square,RMSE and MAE are 0.881,4.529 and3.409 respectively;（3）The improved LUR model was used to simulate the spatial distribution of NO₂/SO₂ concentration.The simulation results show that the high NO₂ concentration area in Anhui province tends to spread from north to south.In spring,areas with high concentration of NO₂ are the least,while in autumn,areas with high concentration of NO₂ are the most,while in Hefei,Maanshan and other economically developed areas,the concentration of NO₂ is relatively higher all year round.The high SO₂ concentration area in Anhui province diffused from northeast to southwest,with the lowest SO2concentration in spring in northern Anhui and summer in southern Anhui.（4）According to the population exposure relative risk model,the spatial division of population NO2/SO2 exposure relative risk in Anhui province is realized.Population NO₂ exposure relative risk is in the order of spring,summer,autumn,winter from small to large according to the seasonal mean value.According to the statistics of cities,the area with the relative exposure risk greater than 1 of Maanshan city is the largest,and which is dominated by medium exposure risk,the overall exposure risk value is relatively high.The exposure risk area of Xuancheng city is the smallest,and the population NO₂ exposure relative risk in the whole city is less than 1.Population SO₂exposure relative risk is in the order of spring,autumn,summer,winter from small to large according to the seasonal mean value.According to the statistics of cities,the area with the relative exposure risk greater than 1 of Fuyang city is the largest,and which is dominated by low exposure risk.The exposure risk area of Xuancheng city is the smallest,and the population SO₂ exposure relative risk in the whole city is less than 1.In summary,the population NO₂/SO₂ exposure relative risk in Xuancheng city is the smallest,and the population health effect is the best.